PART II - INVASIVE PHYSIOLOGICAL ASSESSMENT OF CORONARY DISEASE: NON-HYPERAEMIC INDICES (IFR)
TABLE OF CONTENTS
BOOKSHOP
Updated on May 7, 2021
PART II

Invasive physiological assessment of coronary disease: non-hyperaemic indices (iFR)

Nieves Gonzalo1, Hernán Mejía-Rentería1, Angela McInerney1, Javier Escaned1
1Interventional Cardiology. Hospital Clinico San Carlos, IdISSC, Universidad Complutense, Madrid, Spain

Summary

Non-hyperaemic pressure indices (NHPI), also known as resting pressure indices, constitute a new approach to functional stenosis assessment using intracoronary pressure guidewires. The instantaneous wave free ratio (iFR) was the first NHPI developed to circumvent the dependence of fractional flow reserve (FFR) on maximal coronary hyperaemia, which was associated with patient discomfort and increased procedural costs and time. Upon demonstration of the benefits of iFR, other NHPI have been recently developed. In this chapter we will review the basic principles behind iFR, as well as the validation studies and the results of clinical outcome studies that have established iFR as an evidence-based tool for the assessment of intermediate coronary stenoses recognized in the revascularization guidelines. Data about the use of iFR in special clinical indications such as tandem stenosis, multivessel disease, left main stenosis location, acute coronary syndromes and aortic stenosis will be discussed. Further, we will review the value of iFR to assess disease pattern within a vessel and guide interventions, the co-registration of iFR with coronary angiography, and the assessment of functional PCI results. Other NHPI recently developed will be also discussed.

Introduction

Over the last 20 years the evidence supporting the value of functional assessment of stenosis severity using pressure guidewires culminated in an IA recommendation for clinical use in European clinical practice guidelines [1] . In less than a decade, the use of pressure guidewires for this purpose has been further revolutionised by the introduction of instantaneous wave free ratio (iFR), a simpler, yet as effective index to assess stenosis severity other than fractional flow reserve (FFR). Further developments, facilitated by the fact that iFR is an adenosine-free index, contributed to modifying how intracoronary physiology is used in the catheterisation laboratory, moving from being predominately a method to confirm the indication of revascularisation, to being a valuable way to guide revascularisation and make decisions in complex anatomical subsets. The following sections of this chapter will review the theory behind the use of iFR and other non-hyperaemic pressure indices (NHPI), available evidence supporting their value, and practical aspects regarding their use in different anatomical and clinical settings.

FOCUS BOX 1Background and guideline recommendations for physiological assessment of intermediate stenoses
  • Physiology-based management of coronary artery stenosis has demonstrated improvement in outcomes as compared with angiographic guidance.
  • Fractional flow reserve is a hyperaemic index that requires the use of adenosine.
  • Instantaneous wave free ratio (iFR) is a non-hyperaemic index which evaluates the ratio of distal and aortic pressure during the wave-free period of diastole.
  • Both FFR and iFR have a class IA recommendation for the assessment of the haemodynamic relevance of intermediate stenoses in clinical guidelines of revascularization.

Coronary physiology at rest and during hyperaemia

As discussed in a previous chapter of this book, one of the prerequisites of FFR is that coronary pressure measurements must be obtained during maximal hyperaemia. The cornerstone of FFR is using a proportional relationship between the drop in intracoronary pressure across the stenosis and the decrease in maximal myocardial perfusion caused by that stenosis, and that proportionality is derived from the lineal coronary pressure-flow relationship that only takes place during maximal hyperaemia.

This requirement for adenosine or other potent vasodilator stands behind the quest of non-hyperaemic pressure indices. Around the time iFR was developed it was widely accepted that the low adoption of FFR in clinical practice was largely related to the inconveniences related to the use of adenosine, ranging from patient discomfort to costs and availability of the drug, to disruption of routine established workflow in the catherization laboratory.

The advent of iFR was also facilitated by advances in intracoronary wire and computer technology that enabled on-line, instantaneous measurements of the translesional pressure ratio, allowing the abandonment of the need for averaged coronary pressures that had been used for the calculation of FFR since its inception.

How is it possible to use non-hyperaemic pressure measurements to characterise stenosis severity? The rationale of this alternative approach to FFR can be better understood when we examine the relationship between coronary flow and translesional pressure gradients. The mathematical relationship relating both variables, shown in Figure 1, applies to both hyperaemic and non-hyperaemic indices, and constitutes a true fingerprint of the functional relevance of a coronary stenosis. Figure 1, shows clearly the curvilineal, or non-lineal, character relationship between coronary flow and translesional pressure gradient. This is due to the fact that the contribution of friction (F) and separation or turbulence (S) associated with flow (Q) in generating a translesional pressure gradient (DP) is markedly different, as shown in the equation DP = FQ + SQ2. The functional severity of a coronary stenosis is thus reflected in a characteristic curve of the pressure gradient / flow relationship ( Figure 1 ). The pressure gradient generated by a given stenosis will be part of the same curve either at rest or after increasing myocardial flow by administering adenosine.

It is important at this point to highlight several relevant facts related to the measurement of pressure gradients:

1) The pressure gradient vs flow curve applies to both hyperaemic and non-hyperaemic indices. As a matter of fact, the effect of hyperaemia is to displace the gradient value over the curve until the corresponding hyperaemic flow value ( Figure 1 ).

2) In the coronary arteries the pressure gradient varies over the cardiac cycle. At a difference with a hydraulic analog with constant flow, coronary flow varies significantly over the cardiac cycle. As a result, the pressure gradient waveform mirrors that of coronary flow ( Figure 2 ).

3) The magnitude of a translesional pressure measurements is determined by the time window over which it is obtained. As a legacy of the limited technology used at the time FFR was first proposed (i.e. saline-filled hollow guidewires capable only of transmitting dampened pressure signals), translesional pressure gradient based on averaged pressure measurements obtained over several heart beats are used for both FFR and Pd/Pa. Alternatively, with current wire and computational technology, translesional pressure gradient can be selectively measured within a specific time window and within a single cardiac cycle. ( Figure 3 )

4) Under non-hyperaemic conditions, the microcirculation vasodilates in order to keep resting flow constant ( Figure 4 ).[2]

These facts are important to understand how, in combination with the above-discussed aspects of stenosis physiology, non-hyperaemic pressure measurements can highlight the haemodynamic effect of a stenosis. Increasing coronary flow facilitates differentiating functionally mild and severe stenoses. ( Figure 1 and Figure 3). In that regard, FFR is much better than Pd/Pa as a diagnostic tool. Yet, translesional pressure measurements obtained during the high-flow diastolic interval have a much higher discriminative power than Pd/Pa. Finally, the most powerful index in generating a pressure gradient through a stenosis is the hyperaemic wave free period (WFP) pressure ratio, a selective measurement within diastole associated with the administration of adenosine (an index close to the previously described diastolic FFR) ( Figure 3 ).

iFR development

Fractional flow reserve was developed following a model of coronary physiology in which, for the sake of simplicity, relevant aspects like collateral support and central venous pressure were removed from the initial FFRmyo model. Yet, FFR retained the characteristic of being able to describe, in physiological terms, how a stenosis contributed to decrease myocardial blood flow. So, an FFR value of 0.75 indicated, somehow, that the interrogated stenosis caused a 25% impairment in blood supply to the subtended myocardium. The approach to validate iFR was different. The aim of iFR validation studies was to establish an iFR cut-off value with a similar ability than FFR to classify stenoses in terms of haemodynamic severity.

The diastolic window chosen for iFR interrogation was based on wave intensity analysis. Using this technique, a diastolic window free of waves, which was named the wave free period (WFP) was identified. Within the WFP, which started immediately after the backward decompression wave, there was a cessation of pressure waves originating either from the aorta or the microcirculation. The WFP window was calculated beginning 25% of the way into diastole (identified from the dicrotic notch) and ending 5 ms before the end of diastole (identified from EKG). During the WFP, microcirculatory resistance presented the lowest and more stable values documented over the whole cardiac cycle under resting conditions [3]. Interestingly, microcirculatory resistance values measured at rest within the WFP were virtually identical to those obtained with whole-cycle, averaged measurements of pressure and flow during maximal hyperaemia (in other words, with a whole-cycle averaged approach similar to FFR). iFR was then calculated as the mean pressure distal to the stenosis during the diastolic wave-free period, divided by mean aortic pressure during the diastolic wave-free period.

Thus, in this first approach and in contrast to FFR, calculation of iFR relied on the analysis of both pressure waveforms and EKG signals. Furthermore, as time delay in Pa and Pd based waveforms occurring as a result of different wave of transmission (hydraulic and electric, respectively) might lead to an artefactual diastolic pressure gradient, in iFR both pressure waveforms were aligned at the time of pressure equalisation, to ensure that any difference in pressure over the WFP truly reflected a haemodynamic effect of the interrogated stenosis. Further versions of the iFR algorithm allowed automatic detection of the WFP based solely on the pressure waveform. In the following sections we shall discuss the validation of iFR in different study designs. ( Figure 5 )

iFR clinical and validation studies

The iFR concept has been tested in a number of validation studies with different design, ranging from a direct comparison with FFR, to head-to-head comparison with FFR against non-invasive test of myocardial ischemia and, finally, to non-inferiority randomised clinical trials using FFR as comparator. In the next sections these studies will be analysed in detail.

Studies comparing FFR and iFR

( Table 1 )

The ADVISE (Adenosine Vasodilator Independent Stenosis Evaluation) was the first study assessing the diagnostic performance of iFR against FFR. In a series of 157 coronary stenosis iFR was significantly and closely correlated with FFR (r=0.9, p <0.001). A cut-off value of iFR ≤0.83 was used in this study showing good diagnostic accuracy. The receiver operating characteristics (ROC) area under the curve (AUC) of iFR to predict an FFR<0.8 was 0.93 with a specificity, sensitivity, negative and positive predictive values of 91%, 85%, 85%, and 91%, respectively [3]. The article reporting the results of the ADVISE study included also, for the first time, the rationale and technique of iFR.

Soon after the publication of the ADVISE, a cooperative study (VERIFY) showed that lower cut-off points of iFR (0.80 and 0.83) had lower diagnostic accuracy [4]. These results were discouraging, but soon were overweighed by several new comparisons of iFR with FFR in different clinical scenarios Figure 6. The ADVISE multicentre international registry evaluated the classification agreement of iFR and FFR in 339 intermediate coronary stenosis showing an area under the curve of 0.86 (for an FFR<0.80). The iFR value of 0.89 was determined as the best cut-off value to predict an FFR below 0.80 [5]. A registry in an Asian population with 238 lesions showed a significant iFR-FFR correlation with an r =0.77. The AUC of iFR to predict an FFR<0.80 was 0.9 and the best cut-off value for iFR was ≤0.90 with a sensitivity, specificity, positive and negative predictive values, and diagnostic accuracy of 76%, 86%, 82% and 80%, and 82%, respectively [6].

The ADVISE in Clinical Practice registry assessed 392 intermediate stenosis and demonstrated that iFR maintained a high level of agreement with FFR when measured in clinical practice with commercially available systems (0.87 AUC for FFR<0.80, classification match 80%, and optimal iFR cut-off of ≤0.90 [7].

While the number of studies on iFR was increasing, one of the major problems at that time was that the algorithm used to calculate iFR varied from study to study. In order to provide an independent validation, the investigators of the RESOLVE study pooled data from 1593 lesions included in prior studies at 15 clinical sites. A comparison of iFR and FFR was then performed using a single algorithm in a central core laboratory. Of note, the pooled data included the VERIFY study. The optimal iFR cut point for FFR<0.80 was ≤0.90 with an overall accuracy of 80.4% [8].

One of the problems of a retrospective registry like RESOLVE is that the pressure tracings used in the central core laboratory had been obtained without concomitant EKG, which was used at that time to determine the wave free period. To outline the diagnostic value of the iFR algorithm to be implemented in the first commercially available system, the investigators of the ADVISE II international multicentre study interrogated prospectively 690 lesions with pressure guidewire in an strict, standardised fashion. The results of the study demonstrated a strong correlation between FFR and iFR (r = 0.81, 95% CI: 0.78 to 0.83, p < 0.001). The AUC in ROC analysis was 0.90 (95% CI: 0.88 to 0.92, p < 0.001) and the optimal cut-off value was ≤0.89. Using this cut-off, iFR correctly classified 82.5% of total stenoses (primary endpoint of the study), with a sensitivity of 73.0% and specificity of 87.8%. ADVISE II demonstrated also that the use of a hybrid FFR/iFR algorithm, in which FFR was selectively used to interrogate stenoses with iFR values around its cut-off value, had a very high diagnostic accuracy (94.2% CI: 92.2% to 95.8%)[9]. These findings were subsequently translated to the design of the SYNTAX II trial, the first prospective study using iFR for decision making [10]. The VERIFY II registry evaluated 257 stenosis and used a cut-off value of ≤0.90 for iFR. Diagnostic accuracy versus FFR≤0.80 was calculated with a rate of misclassification of 21% [11]. Similar accuracy (79.9%) of an iFR value of <0.90 to predict and FFR≤0.80 was reported in a registry of 763 patients comparing iFR with FFR and contrast FFR (AUC 0.88) [12]. A small study evaluated the diagnostic performance of iFR in the setting of non-culprit artery stenosis in patients with STEMI. FFR and iFR measurements were obtained at the time of primary PCI and at a staged procedure in 66 non-culprit lesions in 50 patients. ROC analysis showed a high accuracy of iFR to identify an FFR≤0.80 with an AUC of 0.95. For an iFR cut-off value of 0.89 the sensitivity, specificity, positive and negative predictive values were 95%, 90%, 86% and 97%, respectively [13].

In summary, validation studies demonstrated that iFR and FFR values are closely correlated and showed a good diagnostic accuracy of iFR to predict and FFR<0.80.

Studies comparing iFR with other tests of ischemia

( Table 2 )

All previous studies had the common limitation of using FFR as the standard of reference. To circumvent this shortcoming, IFR was validated against different invasive and non-invasive tests of ischemia, in most cases following a head-to-head comparison with FFR. The CLARIFY study evaluated the iFR and FFR agreement with hyperaemic stenosis resistance (HSR), a combined pressure and flow index that calculates the resistance of the stenosis, therefore circumventing limitations of pressure-only indices. Fifty-one stenosis were evaluated, showing similar diagnostic performance of iFR and FFR with HSR giving an AUC of 0.93 for iFR vs. 0.96 for FFR, p = 0.48 [14]. Another study by Van de Hoef et al used HSR and myocardial perfusion scintigraphy as the reference to determine the presence of ischemia. Their results demonstrated similar diagnostic value of iFR and FFR with similar AUC (0.84, and 0.88, respectively; p≥0.20) [15]. Two studies evaluated the diagnostic performance of iFR against PET as the gold standard for ischemia detection. A first study assessed a total of 115 consecutive patients with left anterior descending artery stenosis. The optimal cut-off values and diagnostic performance of FFR and iFR against 13N-ammonia PET-derived coronary flow reserve (CFR) and relative flow reserve (RFR) were tested. iFR and FFR showed similar diagnostic accuracies to predict a CFR <2.0 (FFR 69.6%, iFR 73.9%) and RFR <0.75 (FFR 73.9%, iFR 71.3%) [16]. De Waard et al tested the diagnostic performance of iFR and FFR against [15O]H2 PET hyperaemic myocardial blood flow (MBF) in 320 coronary arteries (136 stenosis). Both indices showed similar AUC (0.78 [95% confidence interval (CI): 0.70–0.85] for FFR, 0.74 (95% CI: 0.66–0.81) for iFR). In coronary stenoses, the diagnostic accuracy compared with impaired PET MBF was 72% (95% CI: 64–79%, К: 0.44) for FFR ≤0.80, 72% (95% CI: 64–80%, К: 0.44) for iFR≤ 0.89 [17]

Finally, one study evaluated the performance of iFR and FFR against invasive coronary flow reserve (CFR) in 216 stenosis and found a significantly stronger correlation and a higher diagnostic performance for iFR (iFR AUC 0.82 versus FFR AUC 0.72; P<0.001, for a CFVR of 2) [18].

In summary, head-to-head comparisons of iFR and FFR against other indices of stenosis severity or myocardial ischaemia demonstrated a similar diagnostic power for both techniques. Interestingly, iFR was found to correlate better with CFR than FFR.

FOCUS BOX 2Validation of iFR
  • Values of iFR and FFR are significantly and closely correlated
  • iFR has a good diagnostic accuracy to predict and FFR<0.80.
  • iFR has been validated against myocardial perfusion scintigraphy and PET showing good diagnostic accuracy with values of area under the curve similar to those of FFR.
  • CFR is more strongly correlated to iFR than to FFR.

Clinical outcomes trials

The ultimate proof to demonstrate the value of iFR in clinical practice was the launching of clinically-oriented, non-inferiority, randomised clinical trials comparing iFR- and FFR-based revascularisation. Two large randomized clinical trials evaluated whether iFR wa a safe and feasible alternative to FFR to guide revascularization. The potential advantage of using iFR instead of FFR would be the avoidance of adenosine thus reducing cost and potential side effects for the patients. Both studies used a single cut-off point of ≤0.89 for iFR.

The DEFINE-FLAIR (Functional lesion assessment of intermediate stenosis to guide revascularisation) was a prospective, multicentre, international, non-inferiority, double blinded trial in which 2492 patients with intermediate severity coronary artery lesions were randomized in a 1:1 ratio to undergo either iFR-guided or FFR-guided coronary revascularization. Both patients with stable angina and non-culprit stenosis in patients with acute coronary syndromes (ACS) were included. The primary endpoint was the 1-year risk of major adverse cardiac events (MACE) which were a composite of death from any cause, nonfatal myocardial infarction, or unplanned revascularization. The non-inferiority margin was 3.4%. The results demonstrated that iFR was not inferior to FFR to guide revascularization with similar outcomes in both groups at 1 year (MACE occurred in 78 of 1148 patients (6.8%) in the iFR group and in 83 of 1182 patients (7.0%) in the FFR group (difference in risk, −0.2 percentage points; 95% confidence interval [CI], −2.3 to 1.8; P<0.001 for noninferiority; hazard ratio, 0.95; 95% CI, 0.68 to 1.33; P = 0.78). No differences were observed in death and in the risk of the individual components of MACE. The study also demonstrated a shorter procedural time (40.5 minutes vs. 45.0 minutes, P = 0.001) and a lower occurrence of adverse symptoms during the procedure (39 patients [3.1%] vs. 385 patients [30.8%], P<0.001) for patients who underwent iFR [19].

The iFR SWEDEHEART (Evaluation of iFR vs FFR in Stable Angina or Acute Coronary Syndromes) had a different study design. It was an open-label multicentre, registry based randomized clinical trial using the Swedish Coronary Angiography and Angioplasty Registry (SCAR) for enrolment. The study population (2037 patients) was similar to DEFINE-FLAIR evaluating intermediate coronary stenosis in stable angina and in non-culprit vessels in ACS that were randomized 1:1 to iFR or FFR guided revascularization. The primary endpoint was similar to the one described for DEFINE-FLAIR (1 year risk of MACE defined as a composite of death from any cause, nonfatal myocardial infarction, or unplanned revascularization) with a noninferiority margin of 3.2%. Concordantly with DEFINE-FLAIR the results confirmed non inferiority of the iFR guided revascularization strategy with an occurrence of the primary end-point event at 1 year in 68 of 1012 patients (6.7%) in the iFR group and in 61 of 1007 (6.1%) in the FFR group (difference in event rates, 0.7 percentage points; 95% confidence interval [CI], −1.5 to 2.8; P = 0.007 for noninferiority; hazard ratio, 1.12; 95% CI, 0.79 to 1.58; P = 0.53). Also in concordance with DEFINE FLAIR, more patients in the FFR group had chest discomfort during the procedure (68.3% vs 3.0% , p<0.001) [20].

A pooled patient level meta-analysis of both trials was performed providing the largest available outcome data for physiology guided revascularization (4529 patients). The two trials evaluated intermediate coronary stenoses and this is reflected in the mean FFR value that was 0.83±0.10 in contrast with previous outcome studies for FFR that showed lower FFR values (0.71 for DEFER and 0.75 for FAME). The pooled analysis demonstrated that a significantly higher proportion of patients were deferred in the iFR-guided than in the FFR guided strategy (1,117 patients (50%) vs 1,013 patients (45%) (p < 0.01)). However, the 1 year MACE rate in the deferred population in the iFR and FFR groups was similar (4.12% vs. 4.05%; HR: 1.13; 95% confidence interval: 0.72 to 1.79; p = 0.60) confirming the non-inferiority of iFR [21]. Reasons for lower deferral with FFR might be related to the higher correlation between iFR and subtended myocardial flow, demonstrated in previous head-to-head comparisons of iFR and FFR against CFR and in studies investigating the causes behind the discrepancy between iFR and FFR in some cases [18] [22]. ( Table 3 )

After the publication of these two pivotal studies, the 2018 guidelines of the European Society of Cardiology on myocardial revascularization gave both FFR and iFR a Class 1A recommendation for guiding revascularization in angiographically intermediate coronary stenoses [1].

FOCUS BOX 3Clinical trial data supporting the use of iFR
  • The DEFINE-FLAIR was a prospective, multicentre, international, double blinded trial in which 2492 patients with intermediate severity coronary artery lesions were randomized 1:1 to undergo either iFR-guided or FFR-guided coronary revascularization
  • The iFR SWEDEHEART was an open-label multicentre, registry based randomized clinical trial in which 2037 patients with intermediate coronary stenoses were randomized 1:1 to iFR or FFR guided revascularization.
  • Both trials demonstrated that iFR was not inferior to FFR to guide revascularization with similar outcomes in both groups at 1 year
  • In both studies there was a lower occurrence of symptoms during the procedure in the iFR group. The DEFINE-FLAIR trial demonstrated also a shorter procedural time with iFR.
  • The pooled analysis of both trials demonstrated that a significantly higher proportion of patients were deferred in the iFR-guided than in the FFR guided strategy but with similar rates of 1 year MACE in the deferred population in both groups.

Furthermore after the publication of the DEFINE FLAIR and iFR SWEDEHERAT trials, a number of relevant pieces of information on the use of iFR in different anatomical and clinical subsets have been published.

Discrepancies between FFR and iFR

Several studies have addressed the possible mechanisms of discordance between FFR and iFR. Cook et al. evaluated baseline and hyperaemic coronary flow velocity and coronary flow reserve (CFR) measured with Doppler in patients with disagreement in stenosis severity classification between iFR and FFR. In FFR+/iFR– discordants, hyperaemic flow velocity and CFR were similar to both FFR–/iFR– and unobstructed groups; In FFR–/iFR+ discordants, hyperaemic flow velocity, and CFR were similar to the FFR+/iFR+ group. These results suggest that i) differences between FFR and iFR can be explained by differences in hyperaemic flow and ii) iFR correlates with coronary hyperaemic flow and CFR better than FFR [22]. The mechanisms of iFR/FFR discrepancy were further tested in a study using PET that included 113 patients. PET derived indices such as myocardial blood flow (MBF), CFR, microvascular resistance, and trans-stenosis resistance were compared between patients with discordant results (FFR≤ 0.80/ iFR≥0.90) and those with concordant abnormal results (FFR≤ 0.80 and iFR<0.90). Resting MBF was similar in the 2 groups but patients with discordant FFR and iFR results showed higher hyperaemic MBF and CFR, compared with the concordant abnormal group (p = 0.003 and p < 0.001, respectively)[30].

Recently, Warisawa et al sought to determine if the pattern of stenosis was a factor influencing discordance between iFR and FFR[31]. Again, most discordance occurred around the cut off points of iFR and FFR. Using pullback assessment, they found that iFR+/FFR- was more frequently found in diffuse pattern of disease while iFR-/FFR+ was more common in focal stenosis. The explanation for these findings is that in the context of diffuse disease, frictional losses along the whole length of the vessel would be the predominant mechanism of pressure loss with minimal change during hyperaemia resulting in an iFR+/FFR- while in the context of focal stenosis, a large change in pressure loss could be expected during hyperaemia resulting in iFR-/FFR+.

Lee et al evaluated the differences in clinical and angiographic characteristics in patients with iFR and FFR discordance. This substudy from the 3V FFR-FRIENDS study evaluated 975 vessels (393 patients) with available pre-intervention FFR and iFR. In a multivariate analysis the authors found that female sex, diabetes mellitus, smaller reference vessel diameter, and higher %DS were significant predictors of low iFR and high FFR. Some of these factors such as diabetes and female sex are also associated with microvascular disfunction and this could be one of the mechanisms explaining the discordance in this group. Conversely, males, absence of diabetes mellitus and lower %DS were significantly associated with high iFR and low FFR. In this group, in accordance with the previously mentioned studies, the disagreement between the 2 indices might be mediated by the presence of a high CFR (the low prevalence of predictors of microvascular disfunction in this group would support this hypothesis)[32].

All the above reinforces the concept that the discrepancy of iFR and FFR values is largely modulated by differences in subtended myocardial flow. The main question is whether such discrepancy results in different patient outcomes. A substudy of the 3V-FFR-FRIENDS evaluated the influence of discordance between FFR and iFR on clinical outcomes. Patients (n=374) with deferred lesions (n=827) were included and MACE (a composite of cardiac death, myocardial infarction, and ischemia-driven revascularization) at 2 years was assessed. The study showed that discordance between FFR and iFR was not associated with an increased risk of MACE. The only group with an increased risk of MACE at 2 years was the one with abnormal values of both FFR and iFR [33].

The lack of clinical relevance of iFR and FFR disagreement can be explained because most of the disagreements occur in the FFR gray zone (between 0.75 and 0.80). Studies that evaluated FFR against test of myocardial ischemia identified FFR 0.75 as the optimal FFR cut-off and the DEFER study demonstrated the safety of this “ischemic” cut-off for deferral. Later, the FFR≤0.80 “clinical cut-off” was implemented in large clinical trials to improve the negative predictive value of FFR in order to avoid leaving significant lesions untreated. Clinical outcome studies using FFR support the finding that in the FFR gray zone, it is equally safe to defer or treat the stenosis. Most of the disagreements between iFR and FFR fall in the FFR gray zone and it is therefore unlikely that they will have prognostic impact[7].

In summary, discordance between resting and hyperaemic pressure derived indices is the result of a complex interaction of clinical, anatomical and physiological characteristics of the patient and the lesion. All these aspects should be considered in the revascularization decisions in the population with discordant measurements. From an overall perspective, discrepant iFR/FFR values typically occur close to the iFR cut-off value, and do not have a detectable influence on the safety of decision making with any of the two indices.

FOCUS BOX 4Discrepancies between FFR and iFR values
  • Discrepancies between FFR and iFR can be explained by different aspects such as differences in hyperaemic flow and microvascular disfunction.
  • These differences between resting and hyperaemic pressure derived indices are the result of a complex interaction of clinical, anatomical and physiological characteristics of the patient and lesion.
  • However, discordance between FFR and iFR has not been associated with an increased risk of MACE in deferred patients.

Use of iFR in special clinical conditions

IFR in left main coronary artery stenosis

The use of physiology guidance in the revascularization of left main stem stenosis has long been controversial particularly because left main stenosis has often been an exclusion criteria in clinical studies. Evidence supporting the safety of FFR to defer revascularisation in LM typically is based on studies comparing outcomes of deferred revasculartisation patients based on non-ischemic FFR with those undergoing revascularisation on the grounds of ischemic FFR values[23]. Applying the same study design, Warisawa recently examined this issue using iFR to guide revascularization in a multicentre registry study[24]. Patients were included if they had a moderate LMS stenosis (40-70%) and could be included if the stenosis extended to the bifurcation and into the ostial LAD or LCx arteries. The iFR reading was taken either just distal to the stenosis in the LMS or, in the presence of ostial LAD or LCx disease, iFR testing in both of these vessels could be performed using the lower iFR value obtained for decision making. The outcome assessed was MACE (composite of death, non-fatal MI, ischaemia driven target lesion revascularization). From a total of 314 patients, 163 were deferred and 151 were revascularized (using either CABG (43.7%) or PCI (56.3%)). No differences were observed in KM survival estimates at 4 years between groups 90.8% versus 85.4% in the deferred versus revascularized groups respectively (HR 1.45, 95% CI 0.75-2.81, p=0.26), nor were there differences in the individual components of the primary end point. Although better outcomes were not seen in the deferral group, these results are consistent with other observational studies using FFR to determine revascularization decisions in the context of left main stenosis. However, no evidence of increased harm was noted using iFR to defer revascularization in this cohort suggesting decision making based on iFR is safe.

Physiology based deferral of LAD stenosis

A recently published analysis of DEFINE FLAIR has addressed the issue about the safety of physiology-based deferral in stenosis located in the LAD. A previous study using FFR as the reference standard had suggested that resting indices were less accurate in evaluating stenosis severity in the LAD than in other vessels with higher incidence of false negatives in this territory [23]. However, the LAD analysis of DEFINE-FLAIR including 872 deferred LAD lesions (421 guided by FFR, 451 guided by iFR) demonstrated a significantly lower event rate when iFR was used in comparison with FFR (2.44% vs. 5.26%; adjusted HR: 0.46; 95% confidence interval [CI]: 0.22 to 0.95; p=0.04). This difference was mainly driven by the lower incidence of unplanned revascularizations in LAD lesions deferred with iFR (2.22% iFR vs. 4.99% FFR; adjusted HR: 0.44; 95% CI: 0.21 to 0.93; p=0.03)[24]. This could be explained by the discordance between FFR and CFR in a subgroup of LAD stenosis. Patients with normal FFR but abnormal CFR have a higher event rate [25]. Additionally, as it has been previously shown, the agreement between FFR and CFR is lower than that between iFR and CFR. Therefore, a discordant result with CFR is more likely when FFR is used to assess a stenosis.

Outcomes in diabetic patients undergoing physiology guided revascularization

Diabetic patients are a well know subgroup of high risk patients. In particular, diabetic patients have a greater risk of restenosis after PCI and have poorer outcomes in terms of MACE in many large clinical trials[29]. A subgroup analysis of the DEFINE FLAIR trial aimed to determine the outcome of diabetic patients undergoing physiology guided revascularization, again using a composite MACE end point of all cause death, non-fatal MI and unplanned revascularization at one year. Seven-hundred and fifty eight diabetics were compared to 1707 non-diabetics. Significantly higher rates of MACE were seen among the diabetic population in comparison to the non-diabetic population (8.5% versus 5.6% respectively, adjusted hazard ratio 1.88 (95% CI 1.28-2.64, p<0.001). Among diabetic patients specifically, no difference was found in MACE rates when revascularization was guided by iFR versus FFR (10.0% versus 7.2%, adjusted HR 1.33 (95% CI 0.78-2.25), p=0.30) and no difference was found in those deferred based on iFR or FFR assessment (MACE in the deferred population (6.8% vs 5.1%; adjusted HR 0.98; (95% CI, 0.38-2.55); p=0.58 confirming that iFR is safe to use as a physiological index in this population.

iFR in multivessel disease

One of the main benefits of guiding coronary revascularization using functional stenoses assessment in the presence of multivessel disease (MVD) is the opportunity to reclassify stenosis severity and to improve decision-making on revascularization strategies. The FAME trial demonstrated that routine FFR-guided PCI in patients with MVD was safer than angiography-guided PCI in terms of MACE [34]. Recently, the two largest randomized trials on coronary physiology in which more than 35% of population had MVD, demonstrated that coronary revascularization guided by the pressure resting index iFR was as safe as FFR regarding clinical outcomes at 1-year of follow-up[19] [20]. Importantly, by removing the need for hyperaemic agents, iFR simplifies functional stenosis assessment in terms of time consumption, adverse side effects and costs, which is especially attractive in the setting of MVD[19]. Data from the iFR-SWEDEHEART trial showed a significantly higher number of lesions evaluated per patient in the iFR arm compared with the FFR arm (1.55±0.86 vs. 1.43±0.70; p=0.002)[20]. Despite the fact that adenosine is safe and well tolerated in most patients, its adverse effects are not negligible and occasionally it can result in a short-lived hemodynamic instability, potentially limiting the possibility of multiple vessels evaluation. Furthermore, in both DEFINE-FLAIR and iFR-SWEDEHEART trials there was a strong trend towards a lower number of stents implanted per patient when PCI was guided by iFR as compared with FFR (DEFINE FLAIR: 0.66±0.92 vs. 0.72±0.96, p = 0.09; iFR-SWEDEHEART: 1.58±1.08 vs. 1.73±1.19, p = 0.05). Therefore, iFR not only avoids adenosine-related discomfort and saves time, but can also potentially reduce costs, especially in MVD clinical subset while remaining as safe as FFR in guiding revascularization.

The second advantage of iFR when assessing MVD relies on the opportunity to change the revascularization strategy. The value of coronary artery bypass grafting (CABG) on hard clinical outcomes over percutaneous intervention has been historically proven by angiographic assessment of stenosis severity without the use of intracoronary physiology[35] [36]. However, it is possible that many native coronary vessels surgically grafted may not have had truly ischemic lesions, and therefore may not have received a true benefit from bypass surgery. Furthermore, in many cases graft occlusion has an asymptomatic clinical course if the native vessel has no-flow limiting stenosis[37]. In this regard, physiological assessment can transform angiographic three-vessel severe disease into one or two vessels with physiologically severe stenoses, potentially modifying the therapeutic strategy. In the FFR-R3F registry (Registre Français de la FFR), changes in therapeutic strategies (medical therapy vs. revascularization either by PCI or CABG) were investigated in 1075 patients that underwent diagnostic angiography with FFR assessment. The authors reported changes of the revascularization decision in about half of the patients after FFR evaluation in comparison with the initial angiography-based strategies. Interestingly the final FFR-based strategies that disagreed with the angiography-based a priori decision were similarly safe at 1 year of follow-up compared with the patients in whom FFR-based strategies matched initial angiography-based decisions[38]. Recently, the SYNTAX II trial demonstrated the value of iFR to guide coronary revascularization in patients with MVD[10]. In this multicentre, all-comers, single arm study, 454 patients with de novo three vessel coronary artery disease with equipoise 4-year mortality risk between a percutaneous or surgical approach underwent multivessel iFR/FFR assessment using a hybrid approach and IVUS guided optimisation of PCI in case of physiologically significant disease. The outcomes were compared with the historical cohort of the original SYNTAX I trial. The main result of this study was that a contemporary state-of-the-art PCI strategy was associated with improved clinical outcomes at one year compared to the predefined PCI arm of the SYNTAX-I study. Importantly, out of 1559 lesions initially intended to be treated according to angiographic appearance, only 75% were found to be functionally significant. In the majority of target lesions (77%) the physiology-based revascularization was only guided by iFR, whereas in the remaining cases (23% of lesions with iFR values between 0.86-0.93) the decision was made based on FFR. Of note, after physiological evaluation of the three coronary vessels only 37.2% of the patients required 3 vessels interventions. Physiological guidance decreased the initial angiography-based 3.5 target lesions per patient to 2.6 lesions per patient, and PCI was deferred in 25% of the lesions. In addition, it has been recently shown that the benefit of the SYNTAX-II strategy is maintained at 2 years of follow-up, with repeat revascularisation occurring in only 1% of deferred lesions[39]. As iFR was used as the sole technique for deciding whether to treat or defer a stenosis in around 75% of lesions in SYNTAX II, this evidence supports the value of iFR to guide clinical-decision making pertaining to revascularization strategies in patients with MVD. The extremely low incidence of repeat revascularisation in lesions with deferred revascularisation at 2 years follow-up in the SYNTAX II trial is reassuring regarding the use of iFR in the high risk anatomical subset of three vessel disease[39].

FOCUS BOX 5iFR in multivessel disease
  • iFR simplifies functional stenoses assessment in terms of time consumption, adverse side effects and costs, aspects especially relevant in the setting of MVD.
  • The SYNTAX II trial have demonstrated the value of iFR to guide coronary revascularization in patients with MVD.
  • A contemporary PCI approach including physiologic evaluation, second generation drug eluting stents and imaging optimization demonstrated improvement of outcomes at 1 year as compared with the PCI cohort of the SYNTAX I trial.
  • In the SYNTAX II trial physiological guidance decreased the initial angiography-based 3.5 target lesions per patient to 2.6 lesions per patient, and PCI was deferred in 25% of the lesions.

Physiological assessment of non-infarct-related arteries in acute coronary syndromes

The presence of MVD in patients with acute myocardial infarction (AMI) is frequent and entails a worse prognosis[40]. There is growing evidence supporting the clinical benefit of the revascularization of non-infarct-related arteries (non-IRA) in AMI patients with MVD after successful treatment of the culprit artery. This strategy was first tested by angiography-guided revascularization in patients with ST-segment elevation myocardial infarction (STEMI), with immediate PCI of non-IRA after primary PCI of the infarct vessel in the PRAMI trial (Preventive Angioplasty in Acute Myocardial Infarction) and by in-hospital complete revascularization in the CvLPRIT trial (Complete versus Lesion-only Primary PCI trial)[41, 42]. Both randomized studies demonstrated a significant reduction on hard clinical outcomes in the complete revascularization arm as compared with the IRA-only revascularization arm. Recently, a complete physiology-guided revascularization strategy using FFR in non-IRA in STEMI patients has been tested. The DANAMI-3—PRIMULTI trial demonstrated a reduction of repeat revascularization at long-term follow up in the FFR-guided complete revascularization arm of non-IRA as compared with no further revascularization after primary PCI of the culprit vessel; however, all-cause death and non-fatal reinfarction were similar between groups[43]. The results of this study have been reproduced by the recent COMPARE-ACUTE study, a randomized trial in STEMI patients in which an FFR-guided revascularization of non-IRA during the index hospitalization was superior to no further revascularization of non-IRA, mainly due to the avoidance of subsequent revascularisations at long term follow-up[44].

Despite the encouraging results of the aforementioned trials, evidence supporting a physiological-guided complete revascularization strategy in STEMI patients remains scarce. In general, data supporting the use of FFR and new resting pressure-based indices like iFR is limited in this clinical scenario, and there are no studies comparing angiography-based revascularization vs. physiology-based revascularization of non-IRA during acute or subacute phase of STEMI. Furthermore, a recent study based on the metanalysis of DEFINE-FLAIR and iFR SWEDEHEART suggests an excess of events at one-year in patients with ACS in whom deferral of revascularisation of non-IRA was based on physiology as compared to patients with stable angina (SA)[21]. This analysis included physiological evaluation of non-IRA in 440 patients presenting with ACS. Overall, ACS as a clinical presentation was associated with a higher MACE rate compared with SA in deferred patients (5.91% vs. 3.64% in ACS and SAP, respectively; fully adjusted hazard ratio: 0.61 in favour of SAP; 95% confidence interval: 0.38 to 0.99; p= 0.04). Interestingly, in the FFR group, deferred patients with ACS had more events than deferred patients with stable angina (HR: 0.52; 95% CI:0.27 to 1; p<0.05) while in the iFR group there were no differences in outcome in deferred patients depending on the clinical presentation (HR:0.74; 95%CI:0.38 to 1.43 p=0.37)[19]. The potential advantages of one of the indices in the evaluation of non-culprit stenoses in patients with ACS deserves further research. The currently recruiting iMODERN trial may go some way to addressing this question by comparing iFR guided revascularization at the time of STEMI versus a delayed perfusion cardiac MRI guided strategy (NCT03298659).

In the AMI clinical subset, factors related to time consumption, technical issues, contrast-associated nephropathy and hemodynamically fragile conditions raise concerns regarding what is the optimal timing (during the index procedure, at another time before discharge, or afterwards) and what is the best method to guide a complete revascularization of non-IRA. The concerns about equivocal decision-making with FFR and iFR in the context of AMI stems from theoretical reasons and clinical evidence. Given the completely different physiological framework in AMI patients as compared with SA patients, the widely demonstrated safety of FFR or iFR in patients with SA cannot be easily transferred to AMI patients. Recently, Thim et al. compared acute iFR values in non-IRA (obtained immediately after successful primary PCI of the culprit vessel) with repeated iFR measurements obtained at a staged follow-up procedure in 120 STEMI patients (157 non-IRA)[45]. Through a median time of 16 days (IQR: 5 to 32 days) between acute and follow-up assessment, authors found a classification agreement of 78% between acute and follow-up iFR values in non-IRA, with acute iFR values in general being lower than follow-up iFR values. However, the overall negative predictive value of acute iFR was high (89%), which means that one of the main benefits of the resting iFR index in the acute clinical subset of STEMI patients would be ruling out ischemic-causing lesions. From a physiological framework, the demonstrated increase in resting coronary flow that occurs in both IRA and non-IRA in AMI patients may partially explains the lower, acute iFR values compared with staged iFR repeated measurements[46]. However, a recent study by Choi et al. is in contradiction with these findings. These authors evaluated FFR, iFR, CFR and IMR in non-IRA from 100 AMI patients and performed a comparative analysis with 203 SA patients across different groups of angiographic severity (i.e. 40% to 80% diameter stenosis). Despite the fact that non-IRA had lower CFR values than SA patients, the authors found that FFR and iFR values were not significantly different between non-IRA and SA-target vessels across all %DS groups, and that there was no significant interaction between the type of clinical presentation and the changes observed[47].

The results of the few studies carried out heretofore on the clinical benefit of pressure-wire measurements in AMI are encouraging. However, there is a need for randomized studies aimed at specifically evaluating the best strategy (guided only by angiography, by FFR or iFR) and the ideal time to perform the physiological assessment of the non-IRA in AMI patients.

FOCUS BOX 6iFR in non-infarct related arteries in acute coronary syndromes
  • Some studies have shown benefit of non-IRA physiologic evaluation to reduce revascularizations at follow up. However, there are no studies comparing angiographic vs physiology-based management of the non-IRA in acute coronary syndromes.
  • There is no clear evidence about when is the optimal time to perform the physiological assessment of the non-IRA in patients with ACS.
  • iFR has a high negative predictive value in the evaluation of non-IRA in the setting of ACS. The increase in resting coronary flow in ACS could decrease the iFR values obtained in the acute setting.
  • Data from randomized clinical trials suggests an excess of events at one-year follow up in patients with acute coronary syndromes in whom revascularization deferral of non-IRA was based on physiology as compared to patients with stable angina. This was more relevant for FFR than for iFR.
  • The potential advantages of one of the indices (FFR or iFR) in the evaluation of non-culprit stenoses in patients with ACS deserves further research.

Physiological assessment of concomitant coronary artery disease in patients with severe aortic stenosis

The concomitant presence of coronary artery disease in patients with symptomatic severe aortic valve stenosis (SAS) is highly frequent and represents a challenge for clinical decision-making and therapeutic strategies[48]. Current clinical guidelines recommend simultaneous revascularization of severe coronary lesions involving proximal segments of main branches in patients with clinical indication for aortic valve treatment[1]. However, such coronary stenosis severity assessment is based on angiographic appearance, despite the well-known inaccuracy of the method in determining the true functional stenosis relevance, even in the presence of SAS[49]. On the other hand, the singular pathophysiological framework triggered by the aortic valve disease, which includes left ventricle hypertrophy, increased afterload and microvascular dysfunction, generates uncertainty about the possibility of translating the well demonstrated accuracy and clinical benefit of pressure-wire-guided coronary revascularization in patients with aortic stenosis[50]. Wiegerinck et al. evaluated the effect of aortic valve stenosis on coronary haemodynamics demonstrating a decreased coronary flow reserve (CFR) and an increased hyperaemic microvascular resistance in the presence of aortic stenosis, and an immediate improvement in the CFR and decrease in the hyperaemic microvascular resistance after transcatheter aortic valve replacement (TAVR)[51]. Despite these changes in coronary haemodynamics caused by the aortic valve disease, several efforts have been made in evaluating the feasibility and reliability of using trans-stenotic pressure gradient indices in assessing the functional relevance of concomitant coronary lesions in SAS patients. Recent studies have demonstrated an overall good reproducibility of FFR when comparing values obtained before and after TAVR, and a good overall diagnostic performance of FFR compared with myocardial perfusion scintigraphy[52] [53]. However, the available evidence of FFR in SAS patients is still scarce, and the hemodynamic adverse effects derived from adenosine use can potentially be more pronounced in this fragile population. In this regard, adenosine-free resting indices are very attractive to overcome the aforementioned limitations of FFR and have the potential to increase the use of invasive physiology to guide coronary revascularization in SAS patients. Recently, Scarsini et al. evaluated the effect of TAVR on iFR values[54]. Although a significant individual iFR variation was found in post with respect to pre-TAVR iFR values, authors reported an overall high diagnostic performance of both iFR pre-TAVR and iFR post-TAVI using FFR post-TAVR as the reference standard (AUC 0.90 [0.84-0.96] pre- and 0.93 [0.88-0.97] post-TAVR). In addition, by using the iFR cut-off ≤0.89, the authors found a very high negative predictive value of both iFR pre-TAVR and iFR post-TAVR in ruling out ischemic-lesions as determined by FFR ≤0.80 (98.9% and 97.5%, respectively), whereas positive predictive value was modest (44.1% and 60.1%, respectively). This last finding could be related to the intrinsic hemodynamic changes caused by the aortic valve disease, such as a compensatory increased baseline flow and low CFR that can decrease the positive predictive value of pre-TAVR iFR. Yamanaka et al. evaluated the diagnostic performance of iFR in 95 SAS patients with intermediate coronary lesions[53]. The authors found an optimal iFR cut-off value of 0.82 to demonstrate myocardial ischemia using both FFR<0.75 and myocardial perfusion scintigraphy as the reference methods, with a high diagnostic performance for iFR (AUC 0.89 and 0.84, respectively). In addition, Ahmad et al. evaluated the effect of the aortic valve disease on coronary pressure, flow, iFR and FFR values obtaining Doppler-derived flow and pressure intracoronary measurements at rest and during hyperaemia, immediately before and after TAVR[55]. The authors found that iFR post-TAVR did not change (0.88±0.09 pre- vs. 0.88±0.09 post-TAVI, p=0.73), whereas FFR post-TAVR significantly decreased (0.87±0.08 pre- vs. 0.85±0.09 post-TAVR; p=0.001). These findings were further confirmed in a small study by Vendrik et al of thirteen patients who underwent pre-TAVI, post TAVI and 6-month physiological assessment of intermediate stenosis[56]. Again, FFR was found to be significantly lower post TAVI both immediately and at 6 month follow up while iFR did not significantly change. These results indicate that FFR could underestimate the severity of coronary stenoses in patients with SAS. The results of these studies are opposed to those reported by Scarsini et al. in which iFR showed a non-homogeneous behaviour after TAVR. Of note, in this latter study a significant proportion of iFR measurements that crossed the iFR threshold after TAVR had pre-TAVR iFR values very close to the cutoff 0.89, which certainly increases the probability of variations[54].

In summary, the few invasive physiological studies carried out to date in patients with SAS have shown that pressure-wire-based assessment of coronary stenoses, with both FFR and iFR, is feasible and technically safe in this clinical subset. However, the underlying pathophysiological framework of aortic stenosis requires caution in interpreting invasive physiological measurements before aortic stenosis treatment. At the time of writing this chapter, outcomes studies assessing the long-term clinical benefit of pressure-wire-guided coronary revascularization specifically in aortic stenosis patients are pending.

FOCUS BOX 7Physiological assessment of coronary artery disease in patients with aortic stenosis
  • Pressure-wire-based assessment of coronary stenoses, with both FFR and iFR, is feasible and technically safe in patients with severe aortic stenosis
  • Resting indices could potentially offer the advantage of avoiding adenosine and its hemodynamic adverse effects in this fragile population.
  • The singular underlying pathophysiological framework of aortic stenosis requires caution in interpreting coronary invasive physiological measurements before aortic stenosis treatment

Sex differences in physiology based revascularization

Kim et al recently examined the differences in outcomes between men and women included in the DEFINE FLAIR study. In this subanalysis, 601 women and 1891 men were included with the primary end point being MACE (a composite of all-cause death, non-fatal myocardial infarction or unplanned revascularization)[28]. There was an even distribution of stenosis location across both groups with the LAD being the most commonly interrogated vessel. No differences in mean iFR values across the sexes were noted, however, FFR values were statistically lower in men than in women. Women had a lower number of functionally significant lesions per patient (0.31 ± 0.51 in women versus 0.43 ± 0.59 in men p<0.001) and were therefore overall less likely to undergo revascularization (42.1% versus 53.1% for women and men respectively, p<0.001). Revascularization rates in women did not differ depending on which physiological index was used (iFr or FFR) while men who had FFR guided procedures more commonly underwent subsequent revascularization in keeping with their lower FFR values. Despite these differences, there was no difference in 1-year MACE between men and women in the overall population (5.49% versus 6.77% for women versus men, adjusted hazard ratio 0.82, 95% CI 0.53-1.28, p=0.380). Similar to the overall findings of DEFINE FLAIR, both iFR and FFR guided revascularization strategies had similar 1-year MACE when stratified by sex.

Value of iFR in performing whole-vessel interrogation and assessing serial stenoses

Physiology-guided revascularization and PCI strategies become challenging in the presence of tandem lesions or diffuse coronary artery disease. Fractional flow reserve was validated for isolated stenosis, and its real value for a given lesion can be masked by the presence of downstream serial stenoses[57]. In other words, FFR assessment tends to underestimate the real contribution of each stenosis to the ischemic burden in the presence of tandem disease. In this subset, the fluid dynamic interaction under hyperaemic conditions between two serial stenoses alters the true value of FFR for any of the two lesions: if the sensor of the pressure-wire is placed between both stenoses, the distal stenosis attenuates the pressure-drop of the circuit in a way that decreases the pressure gradient and underestimates FFR for the proximal lesion. If the sensor of the pressure-wire is placed distally to both stenoses, the obtained FFR value represents the sum of both serial stenoses, and the distal lesion cannot be separately evaluated. To overcome this FFR limitation, De Bruyne et al. developed a mathematical equation to predict the FFR value of each stenosis separately as if the other stenosis were absent[57]. However, as such approach was non-operative due to its complexity (requiring coronary wedge pressure measurements, and not being valid if coronary branches exist between tandem stenoses) it never became adopted in clinical practice. In clinical practice, most operators performed a careful pullback FFR assessment to determine the most relevant stenosis by visual detection of the largest pressure drop and, after treating the most severe stenosis, they repeat the FFR measurement to determine the true relevance of the residual stenosis.

Non-hyperaemic measurements offer a unique opportunity to perform separate analysis of stenosis severity in this context. While under hyperaemic conditions coronary flow decreases and becomes unpredictable as it passes through a ≥40% diameter stenosis, in non-hyperaemic conditions the coronary flow remains relatively constant and stable, regardless of the severity of coronary artery stenosis, due to the compensatory effect of the microcirculation Figure 4 [2, 58]. As a consequence, after removal of a given epicardial stenosis, the coronary flow remains more stable under resting than under hyperaemic conditions, which theoretically means that trans-stenotic pressure changes across serial stenoses become more predictable, leaving pressure gradient in residual lesions largely unchanged[59]. Based on this, iFR can be more accurate than FFR to predict post-PCI physiological outcome, which becomes very attractive in the setting of tandem stenoses or diffuse disease.

The first study that validated iFR in the setting of serial stenoses developed an offline automated physiological map derived from continuous beat-by-beat iFR measurement during resting pullback of pressure wire[60]. The physiological map was used to determine the iFR pressure drop at every millimetre (∆iFR/mm) along the interrogated vessel, to quantify the contribution of each individual stenosis, and to predict the post-PCI iFR outcome after stenting a given stenosis. Authors reported a high accuracy of the pre-PCI physiology map to predict post-PCI iFR outcome, with a mean difference between predicted iFR and actual observed post-PCI iFR of 0.016 ± 0.004.

Co-registration of iFR and angiography to guide coronary interventions

Currently, the more advanced iFR-pullback system (iFR angiography co-registration, Philips BV, The Netherlands) plot, align and overlay the physiological map onto the angiogram using the time-stamp data recorded during a manual pullback, in order to provide a rapid visualization of areas of iFR loss and an easy evaluation of which lesions cause the highest physiological impact on the ischemic burden in the presence of tandem or diffuse disease. This physiological map permits an accurate assessment of the relevance of each individual lesion in tandem stenoses, and the pattern of iFR loss allows the operator to determine whether the disease is predominantly focal or diffuse ( Figure 7 and Figure 8 ). In addition, the software facilitates PCI strategies by allowing operators to perform virtual PCI: once the more physiologically relevant stenosis has been identified using the physiological map, it can be manually selected for virtual removal in order to predict the best possible post-PCI physiological result, and the functional relevance of residual disease. In addition, by integrating the variations in apparent length of the radiopaque pressure guidewire tip (which result from vessel foreshortening), the algorithm behind iFR angiography co-registration allows an accurate assessment of lesion length and virtual stent length. Of note, the accuracy of the system persists with manual pullback, avoiding the need for a cumbersome mechanical pullback device. ( Figure 9 )

Overall, the relevance of all this for PCI guidance are notable, as prediction of the final haemodynamic effect can be performed before stent implantation. The online feasibility of this approach (virtual PCI) has been recently validated in the iFR GRADIENT Registry (Single instantaneous wave-Free Ratio Pullback Pre-Angioplasty Predicts Hemodynamic Outcome Without Wedge Pressure in Human Coronary Artery Disease)[61]. In this prospective multicentre study, Kikuta et al. reported a high accuracy of iFR-pullback measurements to predict post-PCI physiological outcomes in patients with angiographically intermediate tandem or diffuse disease: the expected iFR value after stenting a given stenosis predicted the actual post-PCI iFR result with 1.4 ± 0.5% error in paired pre- and post-PCI iFR measurements performed in 128 patients (134 vessels). In addition, online iFR-pullback changed operator procedural planning in 31% of cases as compared with initial angiography-based strategies with a reduction in lesion number (-0.18 ± 0.05 lesion/vessel; p = 0.0001) and length (-4.4 ± 1.0 mm/vessel; p < 0.0001). Algorithmic interpretation (AI) of pressure wire pullbacks has recently been shown to be non-inferior to human interpretation and may greatly assist in the application of this technology by operators who may not be as familiar with the interpretation of iFR physiological pullbacks and virtual PCI[62]. Since post-PCI physiological result is a strong predictor of long-term clinical outcomes, the online tools provided by iFR-pullback system are very attractive in the setting of tandem or diffuse disease and may facilitate decision-making regarding PCI strategies[63, 64, 65]. However, randomized trials are required to assess the impact on clinical outcomes of using iFR-pullback system to guide PCI strategies as compared with using iFR alone or FFR.

Post PCI physiological assessment

More recently, there has been great interest in the assessment of residual ischaemia using pressure wires. Heretofore pressure wire assessment of an intermediate stenosis has mainly been employed to confirm the indication for revascularisation, or to justify deferral of a lesion. Many operators do not assess post PCI physiology for numerous reasons including time, resources and in the context of FFR additional costs and patient discomfort pertaining to repeated adenosine administration. Furthermore ‘optimum’ post PCI iFR and FFR targets have not been clearly defined. However, with the advent of iFR pullback systems capable of locating the site of pressure loss within the vessel, analysis of post PCI physiological results becomes a quick and attractive option. The DEFINE PCI study (Blinded Physiological Assessment of Residual Ischaemia After Successful Angiographic Percutaneous Coronary Intervention) by Jeremias et al was a multicentre prospective registry examining the functional result of PCI in 562 vessels with pressure tracings being assessed independently by a corelab[66]. Despite operator assessed angiographically successful PCI, residual ischaemia was found in 24% of patients based on an iFR cut off of ≤0.89. Using iFR pullback, the residual ischaemia was identified as being focal in >80% of patients and in the vast majority either related to the stented area of just proximal or distal to the stented area. At 1-year follow up presented at Transcatheter Cardiovascular Therapeutics (TCT) in 2020, an iFR cut off point of <0.95 after PCI was found to be associated with increased cardiac death, spontaneous MI and clinically driven target vessel revascularization[67]. This highlights the potential for targeted post PCI optimization based on physiological assessment using iFR pullback. Further studies should aim to confirm a post PCI physiological target that could impact on outcomes and examine the feasibility of achieving this targeted physiological result.

FOCUS BOX 8iFR in the presence of serial stenoses
  • Fractional flow reserve was validated for isolated stenosis, and its real value for a given lesion can be masked by the presence of downstream serial stenoses
  • iFR can be a better tool for the evaluation of serial stenosis given the more stable behaviour of resting flow across the range of stenosis severity as compared with hyperaemic flow.
  • With a pullback of the pressure wire a resting pressure pullback trace can be created in which the hemodynamic significance of each individual stenosis can be quantified.
  • In serial lesions, the expected iFR result after treatment of an individual stenosis can also be accurately calculated (virtual PCI)
  • The co-registration of iFR pullback and angiography allows the use of physiology to determine the disease pattern (diffuse vs focal stenosis) and guide treatment.

Other non-hyperaemic indices

After the clinical outcome studies demonstrating non inferiority of iFR to guide revascularization a wide variety of new NHPI indices have been developed and even since the previous iteration of this chapter in 2019, the evidence pertaining to the validation of these indices has exponentially increased. Their main characteristics are specified in Table 4 and Figure 11 are discussed in brief in the following section[68].

Pd/Pa is calculated over the whole cardiac cycle with the final value being an average of 5 cardiac cycles. Several studies have evaluated the agreement between Pd/Pa and iFR. Kobayashi et al prospectively enrolled 763 patients and demonstrated a high and significant correlation between the 2 indices. Pd/Pa showed a good diagnostic accuracy to predict an iFR<0.89 with an AUC of 0.98 and a best cutoff value of Pd/Pa≤0.91. The diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value were 93.0%, 91.4%, 94.4%, 93.3%, and 92.7%, respectively[69]. A post-hoc analysis of the ADVISE II trial showed no significant differences in AUC to predict an FFR≤0.80 between iFR and Pd/Pa (difference in C statistic, 0.00 [95% CI: 0.01 to 0.00], p= 0.350)[70].

Another study including data from the 3V FFR-FRIENDS and the IRIS-FFR studies evaluated the relation between Pd/Pa and iFR in 1,024 vessels. Both Pd/Pa and iFR showed similar associations with anatomic and hemodynamic stenosis severity but the percent change of iFR according to the increase in severity was higher than that of resting Pd/Pa indicating a higher sensitivity of this index. Similarly, both indices were significantly associated with the 2 year risk of MACE but iFR showed a lower maximum difference in estimated MACE rates. This was due to the lower iFR measurement variability in comparison with Pd/Pa[71]. This may be explained by Pd/Pa, being a whole cycle index may be more susceptible to hemodynamic conditions and the obtained value may be less inaccurate in the presence of pressure drift.

dPR, the diastolic pressure ratio, is an index calculated over the whole diastolic period. This index was recently retrospectively validated against iFR and FFR showing a high correlation between iFR and dPR values (R=0.997; P<0.001) although with a lower correlation with FFR (R=0.77, p<0.001)[72]. Using FFR as a reference the optimal cut-off point to predict an FFR of ≤0.80 was 0.91 giving a good diagnostic accuracy (AUC 0.86, 95% CI 0.78-0.93). However, subsequent studies have validadted a cut off of ≤0.89 showing good agreement of dPR and iFR with mutual differences of 0.006 ± 0.011 [73].

RFR or resting whole cycle ratio is an index whose numerical value is the lowest calculated Pd/Pa anywhere in the cardiac cycle. The VALIDATE RFR study published in 2018 was a retrospective study where RFR was determined from 651 pressure waveforms[65]. The optimal cut off for RFR to detect an FFR≤0.80 was ≤0.89. Using this cut-off, correlation with iFR was high (R2=0.99, p<0.001) with a diagnostic accuracy of 97.4%, sensitivity 98.2%, specificity 96.9%, positive predictive value 94.5%, negative predictive value 99.0%, and AUC of 0.996. Interestingly, in this study RFR was calculated outside of diastole in 12.2% of cases. More recently, the same study group prospectively validated RFR against iFR in 501 pressure recordings from 431 patients[74]. Pressure tracings were assessed in a blinded fashion by a corelab. RFR was tested for equivalence with iFR using a prespecified 1% margin. RFR demonstrated equivalence to iFR with a diagnostic accuracy of 97.8%, sensitivity 97.8%, specificity 97.8%, positive predictive value 96.2%, negative predictive value of 98.7% and an AUC of 0.96 (0.94-0.97, p<0.001). While both of these studies demonstrate promise using the RFR index, neither examine clinical outcomes when revascularization decisions are based on RFR.

Lee et al therefore evaluated the agreement between RFR and dPR with iFR and also looked at their correlation with myocardial ischaemia as defined by PET imaging and finally assessed their association with outcomes. In 1,024 vessels (435 patients), RFR and dPR were significantly correlated with iFR and this correlation was higher than with FFR. Both indices showed a high agreement with iFR (c-index 0.987 and 0.993). In the subgroup assessing correlation with myocardial ischaemia, 115 patients who had also PET imaging were included and no differences in diagnostic performance for the prediction of PET determined myocardial ischemia between iFR, RFR, and dPR were found. A composite outcome of cardiac death, vessel-related myocardial infarction and vessel-related ischemia-driven revascularization in 864 deferred vessels showed that the 3 indices were significantly associated with the risk of events at 2 years (iFR per 0.1 increase: HR 0.514 [95% CI 0.370-0.715], p<0.001; RFR per 0.1 increase: HR 0.524 [95% CI 0.378-0.725], p<0.001; dPR per 0.1 increase: HR 0.587 [95% CI 0.436-0.791], p<0.001)[73]. In a further study by this research group assessing vessel orientated outcomes in deferred lesions, Lee et al found that discordance between resting indices (iFR, RFR and dPR) and FFR did not result in increased VOCO in deferred lesions at 2 years[75]. At 5-years however, concordant normal lesions had the best outcome in terms of VOCO while discordant lesions had worse outcomes than concordant normal lesions but similar to revascularized lesions suggesting the long term safety of deferring revascularization when either one of the FFR or the non-hyperaemic index is negative[76].

Recent interest in the prediction and measurement of post PCI physiology led Omori et al to design and carry out the REFINE-RPG study (Revascularization Based on Non-Hyperemic Pressure Ratio Pullback Guide) which randomized 150 vessels from 140 patients on a 1:1:1 basis to undergo iFR, RFR or dFR pullback assessment[77]. PCI was then performed on physiologically significant lesions (using intracoronary imaging guidance) followed by repeat physiological assessment. The aim was to determine the agreement between post PCI physiological assessment with pre-PCI predicted physiology. A strong correlation between predicted and actual post PCI physiology was found for all three indices (iFR r=0.83 [95% CI 0.72-0.90, p<0.001], RFR r=0.84 [95% CI 0.73-0.91, p<0.001], and dPR r=0.84 [95% CI 0.73-0.91, p<0.001]). Furthermore, across all three indices, no differences were found between the predicted post PCI physiological result and the actual post PCI physiological result (p=0.550). These results show highlight the potential for vessel mapping using RFR and dPR, allowing precision stenting and accurate prediction of results.

The good agreement of these non-hyperaemic indices calculated in diastole with iFR make it possible to predict similar results from the clinical point of view. For whole cycle indices there are some concerns about the impact of haemodynamic conditions, sensitivity to drift and physiological relevance of gradients detected during systole. ( Figure 10 ). While clinical correlations such as the study by Lee et al. are encouraging, they do not replace randomized controlled data such as that available for iFR[72].

Conclusion

The use of coronary physiology to guide revascularization has been found to improve patient outcomes as compared with angiographic assessment. The use of FFR however, had remained low in catheter laboratories worldwide. Among the potential reasons to explain its low adoption could be the need for adenosine that involves cost and increased procedural time together with some potential discomfort for the patient. This has led to the development of resting pressure-based indices capable of providing information about the functional impact of stenosis without the need for inducing hyperaemia. iFR is a non-hyperaemic index which evaluates the ratio of distal and aortic pressure during the wave-free period of diastole. The 2 largest clinical randomized trials using physiology performed until the date have validated iFR as an evidence-based tool for the evaluation of intermediate coronary stenoses. In comparison with FFR, the use of iFR reduced procedural time and patient discomfort without differences in MACE between the groups using FFR or iFR guided revascularization. Data indicates also that the use of this resting index could improve the evaluation of serial stenosis and could determine patterns of disease (focal or diffuse) using pressure pullbacks. This can be relevant to decide not only whether treatment is needed or not but to guide the intervention. In this regard, co-registration of iFR with angiography offers a very powerful tool to generate a physiological map of the vessel and determine the areas amenable to interventional treatment with the additional possibility of predicting the result of treatment. A new era in the use of physiology in the catheterization lab has started with this tool used not only to justify but also to guide interventions and assess their results.

Personal perspective

Pressure-guidewire interrogation has become an indispensable way to improve the outcomes of patients in whom coronary revascularisation is being considered. While the evidence supporting this fact over the last 20 years was largely derived from FFR-based studies, adoption of its use in clinical practice lagged over years. The arrival of iFR, less than 10 years ago, truly revolutionised the field of physiological assessment of coronary stenoses. Once initial reservations on whether iFR might be non-inferior to FFR were cleared by the results of large randomised trials, it became evident that the use of a non-hyperaemic index (NHPI) had many things to offer to the interventionalist and the patients, even beyond the very important aspects of reduced discomfort, lower costs and faster procedures. As dedicated software for iFR analysis was developed, studies demonstrated the improved capabilities of iFR for longitudinal vessel analysis and multiple stenoses, prediction of haemodynamic result of the intervention, and co-registration of iFR with angiography, the latter a true game-changer that virtually transformed physiology in an imaging technique. The best proof of the value of these developments are the many NHPI developed and commercialised over the last three years, including DPR, RFR and dFR. While further testing of these new NHPI will be needed, the excellent correlation with iFR values demonstrated in the so-far limited number of studies available anticipates that many of them will be surely used in the catheterisation laboratory along iFR and FFR. Today, the interventional cardiologist has a much richer toolbox for making decisions on functional stenosis severity in different scenarios by using both hyperaemic and NHPI.

SHARE YOUR COMMENT